WizardLM-2
About
The WizardLM-2 is a family of advanced large language models (LLMs) developed by Microsoft AI. This series includes three cutting-edge models: WizardLM-2 8x22B, WizardLM-2 70B, and WizardLM-2 7B. The flagship model, WizardLM-2 8x22B, is a Mixture of Experts (MoE) architecture with 141 billion parameters, built upon the Mixtral-8x22B-v0.1 base model. These models demonstrate highly competitive performance in complex chat, multilingual tasks, reasoning, and agent-based interactions, often rivaling or surpassing proprietary models in benchmarks like MT-Bench. The WizardLM-2 family was trained using a fully AI-powered synthetic training system, which contributes to their advanced capabilities in various domains including writing, coding, mathematics, and multilingual tasks.
Specifications(4 models)
| Model | Released | Context | Parameters | Structured Outputs |
|---|---|---|---|---|
| WizardLM-2 8x22B | 2024-01 | — | 8x22B | Yes |
| WizardLM-2 70B | 2024-01 | — | 70B | No |
| WizardLM-2 7B | 2024-01 | — | 7B | Yes |
| Together AI WizardLM-2-8x22B | 2024-01 | 33K | 176B | Yes |
Available From(5 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| WizardLM-2 7B | DeepInfra | $0.05 | $0.15 | Serverless |
| WizardLM-2 7B | Lepton AI API | $0.07 | $0.07 | Serverless |
| WizardLM-2 8x22B | Lepton AI API | $0.5 | $0.5 | Serverless |
| WizardLM-2 8x22B | OpenRouter | $0.62 | $0.62 | Serverless |
| WizardLM-2 8x22B | DeepInfra | $0.65 | $0.65 | Serverless |
| Together AI WizardLM-2-8x22B | Together AI | $1 | $1.5 | Serverless |
| WizardLM-2 8x22B | OctoAI API | $1.2 | $1.2 | Serverless |
Frequently Asked Questions
- What is WizardLM-2?
- The WizardLM-2 is a family of advanced large language models (LLMs) developed by Microsoft AI. This series includes three cutting-edge models: WizardLM-2 8x22B, WizardLM-2 70B, and WizardLM-2 7B. The flagship model, WizardLM-2 8x22B, is a Mixture of Experts (MoE) architecture with 141 billion parameters, built upon the Mixtral-8x22B-v0.1 base model. These models demonstrate highly competitive performance in complex chat, multilingual tasks, reasoning, and agent-based interactions, often rivaling or surpassing proprietary models in benchmarks like MT-Bench. The WizardLM-2 family was trained using a fully AI-powered synthetic training system, which contributes to their advanced capabilities in various domains including writing, coding, mathematics, and multilingual tasks.
- How many models are in the WizardLM-2 family?
- The WizardLM-2 family contains 4 models.
- What is the latest WizardLM-2 model?
- The latest model is WizardLM-2 8x22B, released in 2024-01.
- How much does WizardLM-2 cost?
- WizardLM-2 models range from $0.05/1M to $1.2/1M input tokens depending on the model and provider.
